Visual Question Answering on InfoVQA (ANLS)
78.98ANLS ScoreInternVL-2.5-26B*
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| InternVL-2.5-26B*Latency=0.74542025.12 | 78.98 | — | — | — | |
| LUVCBackbone=InternVL-2.5-26B*, Latency=0.4302, Speedup=1.732025.12 | 77.55 | — | — | — | |
| VTWBackbone=InternVL-2.5-26B*, Latency=0.4637, Speedup=1.612025.12 | 75.52 | — | — | — | |
| InternVL-2.5-8B*Latency=0.29852025.12 | 75.45 | — | — | — | |
| LUVCBackbone=InternVL-2.5-8B*, Latency=0.1857, Speedup=1.612025.12 | 74.75 | — | — | — | |
| VanillaModel=Qwen2-VL-7B2026.03 | 73.99 | 51.98 | 100 | 907.4 | |
| PinPointModel=Qwen2-VL-7B2026.03 | 71.4 | 28.88 | 55.5 | 555.5 | |
| PACTBackbone=InternVL-2.5-26B*, Latency=0.4690, Speedup=1.592025.12 | 68.46 | — | — | — | |
| VTWBackbone=InternVL-2.5-8B*, Latency=0.1804, Speedup=1.652025.12 | 67.94 | — | — | — | |
| GPT-4o-202408062024.10 | 66 | — | — | — | |
| PACTBackbone=InternVL-2.5-8B*, Latency=0.1980, Speedup=1.512025.12 | 64.67 | — | — | — | |
| Qwen2-VL-2BParams (B)=2.12024.10 | 64.1 | — | — | — | |
| InternVL2-2BParams (B)=2.12024.10 | 57.7 | — | — | — | |
| Random samplingModel=Qwen2-VL-7B2026.03 | 53.45 | 29.37 | 56.5 | 530.4 | |
| FastVModel=Qwen2-VL-7B2026.03 | 51.3 | 31.69 | 61 | 532.7 | |
| Phi-3-Vision - MicrosoftParams (B)=4.22024.10 | 49.9 | — | — | — | |
| PDropModel=Qwen2-VL-7B2026.03 | 45.06 | 32.7 | 62.9 | 500.1 | |
| Cambrian-13BParams (B)=132024.10 | 44.3 | — | — | — | |
| H2OVL-Mississippi-2BParams (B)=2.12024.10 | 43 | — | — | — | |
| MiniCPM-V-2Params (B)=2.82024.10 | 38.2 | — | — | — | |
| PaliGemma-3B-mix-448Params (B)=2.92024.10 | 34.1 | — | — | — | |
| UNITBase LVLM=Llava-Next, Resolution Strategy=high-resolution grid slicing, Resolution=4482024.09 | 31.9 | — | — | — | |
| PinPointModel=LLaVA-NeXT-7B2026.03 | 30.24 | 25.48 | 65.3 | 381.6 | |
| SigLIPBase LVLM=Llava-Next, Resolution Strategy=high-resolution grid slicing, Resolution=3842024.09 | 29.7 | — | — | — | |
| CLIP-LBase LVLM=Llava-Next, Resolution Strategy=high-resolution grid slicing, Resolution=3362024.09 | 29.3 | — | — | — | |
| VanillaModel=LLaVA-NeXT-7B2026.03 | 25.52 | 38.98 | 100 | 569.5 | |
| SparseVLMModel=LLaVA-NeXT-7B2026.03 | 24.28 | 27.45 | 70.4 | 471.1 | |
| Random samplingModel=LLaVA-NeXT-7B2026.03 | 23.87 | 24.62 | 63.2 | 360.7 | |
| PDropModel=LLaVA-NeXT-7B2026.03 | 23.35 | 26 | 66.7 | 392 | |
| FastVModel=LLaVA-NeXT-7B2026.03 | 23.06 | 26.22 | 67.3 | 386 | |
| ToMeModel=LLaVA-NeXT-7B2026.03 | 19.75 | 26.67 | 68.4 | 315.1 |